hardware

Real silicon.
Published datasheet.

Every customer gets the same SKU. No "performance tier" upsell, no mystery instance type names. If you can read a datasheet, you already know what you're getting.

[ 01 / hardware ]
hardware spec

Real silicon.
Not a slide.

Every node is a published datasheet. Same chips, same disks, same network — for everyone.

compute
AMD & Intel x86
arch modern server-class
cores dedicated on Pro+
mix your choice at launch
memory
ECC
error correction ecc throughout
overcommit none on Pro+
consistency uniform fleet-wide
storage
Network NVMe
media all-flash u.2
replication redundant
snapshots scheduled + on-demand
network
50 Gbps / host
per-host uplink 50 gbps
backbone 100 gbe
ipv6 native /64
[ rack / detail ]
rack detail

What's actually in the box.

Compute

Modern server-class x86 — both AMD and Intel SKUs in the fleet so you can match your workload's price/performance curve. Dozens of SKUs across compute-, memory-, and storage-optimized profiles, all billed per second.

arch amd + intel x86
sku catalog compute / memory / storage tuned
billing per-second, monthly cap
virtualization hardware-accelerated
console browser + ssh

Memory

ECC across the fleet. Same memory profile in every host, every region — no surprise downclocking based on which node you happened to land on.

type ecc
overcommit none on Pro+

Storage

All-flash NVMe, network-attached so a host failure doesn't take your data with it. Replicated across drives at the storage tier — drive loss is survivable without intervention. Snapshots are first-class: schedule them with cron expressions and a retention count, restore in a single API call.

media all-flash u.2 nvme
topology network-attached, replicated
resize online (no downtime)
snapshots scheduled + on-demand
shared filesystems rwx for multi-vm workloads

Network

50 Gbps per host on a 100 GbE backbone. Native IPv6 /64 per project, IPv4 included. Per-VM and per-network bandwidth caps when you need them. Traffic between your resources in the same region is free; egress to the internet bills against your plan's transfer allowance, and we don't charge once you're past it without saying so.

host uplink 50 gbps
backbone 100 gbe
ipv6 native /64 per project
bandwidth caps per-vm + per-network
redundancy ha at every layer
[ 04 / waitlist ]
offer first 100 users · 50% off · 3 months

Get on the metal,
at half price.

We're shipping closed beta now. Drop your email — first hundred founders get half off everything for the first three months. No credit card. No "schedule a call."

// waitlist.signup accepting
$ join_waitlist --offer=beta100
spots claimed 0 / 100